Underwater oil spill reconnaissance and delineation is challenging as the spilled oil can cover a large area and may form plumes beneath the water surface. Autonomous underwater vehicles (AUVs), with improved intelligence, are used more widely for oil spill tracking nowadays. Underwater gliders, a type of AUVs, are favorable for underwater oil spill mapping as they can work for longer durations with less energy storage required. This thesis investigates the capabilities of gliders, especially multiple gliders, as platforms for mapping subsurface oil plumes.

Considering the limited payload capability and energy availability of a glider, a lightweight Cyclops submersible fluorometer and a Ping360 sonar were used on gliders to detect oil in the water in this thesis. The sensors were tested in a tank experiment with oil and the cross-validation from fluorescence measurements and sonar images was expected to improve the reliability in detecting oil. Furthermore, a Slocum glider was developed as a platform for the sensors before it was tested in the ocean with air bubbles which were used as proxies for oil droplets to avoid discharging oil into the environment. This glider was also developed with a backseat driver controller to provide it with an intelligence to adaptively map underwater oil plumes. In addition, a cooperation strategy was proposed for multiple gliders to delineate underwater oil patches simultaneously to overcome the challenges in oil spill mapping such as spatiotemporal aliasing and to improve data redundancy. In this cooperation strategy, a scout glider was commanded to follow a lawn-mower path to cover the area and find potential patches with rich information for follower gliders. A data compression method was designed for the scout glider to reduce the amount of information to be transmitted to the follower gliders. An adaptive path planning strategy was proposed for the follower gliders to

ensure they spent most of their mission time inside patches. The proposed cooperation strategy was compared with other strategies without cooperation and/or without adaptive control and the influence of having adaptive control and multiple gliders on a strategy was investigated through simulations. This developed Slocum glider was used as a follower glider to test the adaptive control in my cooperation strategy through a field experiment.

The tank experiment with oil and the field experiment with air bubbles conducted in this thesis proved the feasibility and necessity of having two or more sensors to cross-validate their measurements as a single sensor was not reliable in proving the existence of a particular substance, such as oil droplets, in the water. The developed backseat driver was able to provide the glider with an ability of adaptive control. However, having adaptive control or multiple gliders cannot guarantee a good score of performance when delineating underwater oil patches. The proposed cooperation strategy with multiple gliders was found to have the best score of performance, especially for long-endurance missions. The performance of the cooperation strategy could be further improved by having a thruster for the follower gliders to overcome the influence of the ocean environment, such as strong currents, when mapping moving underwater patches.